[PATCH v2 1/2] struct_union_enum_specifier: always initialize sym->scope

[PATCH v2 1/2] struct_union_enum_specifier: always initialize sym->scope

[Oleg Nesterov]

Currently it is not possible to figure out the scope of the private
struct/union/enum type, its ->scope is NULL because bind_symbol() is
not called.

Change struct_union_enum_specifier() to set sym->scope = block_scope
in this case, this is what bind_symbol() does when type has a name.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
---
 parse.c | 2 ++
 scope.c | 5 +++++
 scope.h | 1 +
 3 files changed, 8 insertions(+)

diff --git a/parse.c b/parse.c
index a08165a..3e7c60e 100644
--- a/parse.c
+++ b/parse.c
@@ -772,6 +772,8 @@ static struct token *struct_union_enum_specifier(enum type type,
 	}
 
 	sym = alloc_symbol(token->pos, type);
+	/* make toplevel(sym->scope) work */
+	set_current_scope(sym);
 	token = parse(token->next, sym);
 	ctx->ctype.base_type = sym;
 	token =  expect(token, '}', "at end of specifier");
diff --git a/scope.c b/scope.c
index cbf2fcf..420c0f5 100644
--- a/scope.c
+++ b/scope.c
@@ -40,6 +40,11 @@ struct scope	*block_scope = &builtin_scope,		// regular automatic variables etc
 		*file_scope = &builtin_scope,		// static
 		*global_scope = &builtin_scope;		// externally visible
 
+void set_current_scope(struct symbol *sym)
+{
+	sym->scope = block_scope;
+}
+
 void bind_scope(struct symbol *sym, struct scope *scope)
 {
 	sym->scope = scope;
diff --git a/scope.h b/scope.h
index d9a14aa..3cad514 100644
--- a/scope.h
+++ b/scope.h
@@ -53,6 +53,7 @@ extern void end_symbol_scope(void);
 extern void start_function_scope(void);
 extern void end_function_scope(void);
 
+extern void set_current_scope(struct symbol *);
 extern void bind_scope(struct symbol *, struct scope *);
 extern void rebind_scope(struct symbol *, struct scope *);
 
-- 
2.5.0

[PATCH v2 2/2] dissect: fix sym_is_local(SYM_STRUCT/UNION/ENUM)

[Oleg Nesterov]

Now that struct_union_enum_specifier() always sets sym->scope we can
simplify sym_is_local(sym) and rely on toplevel() even if sym is type.

Test-case:

	// copied from linux kernel
	# define __force	__attribute__((force))
	#define WRITE_ONCE(x, val) \
	({							\
		union { typeof(x) __val; char __c[1]; } __u =	\
			{ .__val = (__force typeof(x)) (val) }; \
		__write_once_size(&(x), __u.__c, sizeof(x));	\
		__u.__val;					\
	})

	void func(int *p)
	{
		WRITE_ONCE(*p, 0);
	}

before this patch the widely used WRITE_ONCE() generates a lot of spam which
can't be filtered out using sym_is_local(),

	11:6                    def   f func                             void ( ... )
	11:11  func             def . v p                                int *
	13:9                    def   s :__u
	13:9                    --- . v p                                int *
	13:9                    def   m :__u.__val                       int
	13:9                    def   m :__u.__c                         char [1]
	13:9   func             def . v __u                              union :__u
	13:9   func             -w- . v __u                              union :__u
	13:9   func             -w-   m :__u.__val                       int
	13:9   func             --- . v p                                int *
	13:9   func             --r   f __write_once_size                bad type
	13:9   func             -r- . v p                                int *
	13:9   func             -r- . v __u                              union :__u
	13:9   func             m--   m :__u.__c                         char [1]
	13:9   func             --- . v p                                int *
	13:9   func             --- . v __u                              union :__u
	13:9   func             ---   m :__u.__val                       int

plus it triggers warning("no context") in test-dissect.c. With this patch
the only "nonlocal" report is __write_once_size() call.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
---
 dissect.c | 4 ++--
 dissect.h | 3 ++-
 2 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/dissect.c b/dissect.c
index 40baf64..c48214b 100644
--- a/dissect.c
+++ b/dissect.c
@@ -152,7 +152,6 @@ static inline struct symbol *expr_symbol(struct expression *expr)
 		if (!sym) {
 			sym = alloc_symbol(expr->pos, SYM_BAD);
 			bind_symbol(sym, expr->symbol_name, NS_SYMBOL);
-			sym->ctype.modifiers = MOD_EXTERN | MOD_TOPLEVEL;
 			sym->kind = expr->op ?: 'v'; /* see EXPR_CALL */
 		}
 	}
@@ -238,7 +237,8 @@ static void examine_sym_node(struct symbol *node, struct symbol *parent)
 				return;
 
 			dctx = dissect_ctx;
-			dissect_ctx = NULL;
+			if (toplevel(base->scope))
+				dissect_ctx = NULL;
 
 			if (base->ident || deanon(base, name, parent))
 				reporter->r_symdef(base);
diff --git a/dissect.h b/dissect.h
index 326d3dc..a77a932 100644
--- a/dissect.h
+++ b/dissect.h
@@ -4,6 +4,7 @@
 #include <stdio.h>
 #include "parse.h"
 #include "expression.h"
+#include "scope.h"
 
 #define	U_SHIFT		8
 
@@ -29,7 +30,7 @@ extern struct symbol *dissect_ctx;
 
 static inline bool sym_is_local(struct symbol *sym)
 {
-	return sym->kind == 'v' && !(sym->ctype.modifiers & MOD_TOPLEVEL);
+	return !toplevel(sym->scope);
 }
 
 extern void dissect(struct reporter *, struct string_list *);
-- 
2.5.0

Re: [PATCH v2 1/2] struct_union_enum_specifier: always initialize sym->scope

[Luc Van Oostenryck]

On Wed, Feb 19, 2020 at 05:29:11PM +0100, Oleg Nesterov wrote:
> Currently it is not possible to figure out the scope of the private
> struct/union/enum type, its ->scope is NULL because bind_symbol() is
> not called.
> 
> Change struct_union_enum_specifier() to set sym->scope = block_scope
> in this case, this is what bind_symbol() does when type has a name.

Thanks.
I've just changed the comment to "used by dissect" because
elsewhere the scope or toplevel()s only relevant for symbols.

-- Luc

Re: [PATCH v2 1/2] struct_union_enum_specifier: always initialize sym->scope

[Oleg Nesterov]

On 02/20, Luc Van Oostenryck wrote:
>
> On Wed, Feb 19, 2020 at 05:29:11PM +0100, Oleg Nesterov wrote:
> > Currently it is not possible to figure out the scope of the private
> > struct/union/enum type, its ->scope is NULL because bind_symbol() is
> > not called.
> >
> > Change struct_union_enum_specifier() to set sym->scope = block_scope
> > in this case, this is what bind_symbol() does when type has a name.
>
> Thanks.
> I've just changed the comment to "used by dissect"

Great, thanks!

> because
> elsewhere the scope or toplevel()s only relevant for symbols.

Cough... can't resist ;)

Not really, see struct_union_enum_specifier()->is_outer_scope(). But
yes sure, this is only when ->ident != NULL.

Oleg.

Re: [PATCH v2 1/2] struct_union_enum_specifier: always initialize sym->scope

[Luc Van Oostenryck]

On Thu, Feb 20, 2020 at 12:57:37PM +0100, Oleg Nesterov wrote:
> On 02/20, Luc Van Oostenryck wrote:
> >
> > On Wed, Feb 19, 2020 at 05:29:11PM +0100, Oleg Nesterov wrote:
> > > Currently it is not possible to figure out the scope of the private
> > > struct/union/enum type, its ->scope is NULL because bind_symbol() is
> > > not called.
> > >
> > > Change struct_union_enum_specifier() to set sym->scope = block_scope
> > > in this case, this is what bind_symbol() does when type has a name.
> >
> > Thanks.
> > I've just changed the comment to "used by dissect"
> 
> Great, thanks!
> 
> > because
> > elsewhere the scope or toplevel()s only relevant for symbols.
> 
> Cough... can't resist ;)
> 
> Not really, see struct_union_enum_specifier()->is_outer_scope(). But
> yes sure, this is only when ->ident != NULL.

Ah yes, sorry, I wasn't clear enough here. By 'symbol' here, I
effectively meant: "a name (with its associated semantic)", not
a 'struct symbol'.
 
-- Luc